Biogeosciences Discuss., 9, C2463–C2503, 2012
www.biogeosciences-discuss.net/9/C2463/2012/
© Author(s) 2012. This work is distributed under
the Creative Commons Attribute 3.0 License.
Biogeosciences
Discussions
BGD
9, C2463–C2503, 2012
Interactive
Comment
Interactive comment on “Current state and future
scenarios of the global agricultural nitrogen
cycle” by B. L. Bodirsky et al.
B. L. Bodirsky et al.
[email protected]
Received and published: 17 July 2012
1) My first suggestion is to change the world agriculture into cropland, since the
model used does not cover the role of grasslands.
We agree that our model only covers grassland partially (only grazing from grasslands
and excretion of manure on grasslands). However, "cropland" would also fall short of
the actual model, because large parts of the model describe the livestock sector as
well as the up-stream processing and consumption of products.
Missing flows are mainly nitrogen fixation on pastures, and atmospheric deposition on
pastures. We chose deliberately not to cover these flows as they are mostly depending
on the definition of pastureland and therefore of little use.
C2463
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
We would therefore prefer to stick with the current title.
In addition, there are some important questions regarding the approach to compute the nitrogen inputs in crop production systems, and the efficiency of nitrogen use.
2. A second suggestion is to add a discussion of the nitrogen use efficiency
as used in agronomy. There are various definitions. The definition used in this
paper is completely different from what is generally used, and a comparison and
discussion is needed of the advantage of the definition used compared to others.
BGD
9, C2463–C2503, 2012
Interactive
Comment
Based on the suggestion of the reviewer we now use two definitions for nitrogen efficiency in our article:
• Nr uptake efficiency (NUE) as defined by Dawson et al. (2008) as the ratio between Total N in plant and N supply. This definition is used to compare our results
to other studies.
• Soil Nr uptake efficiency (SNUE), defined as the share of the Nr inputs to soils
taken up by the plant. This definition is used within in the model to estimate inorganic fertilizer requirements. The reason for including this new type of definition
is, that nitrogen fixed from the atmosphere by legumes as well as seed are not
subject to losses prior to uptake. (In this context, please also have a look also on
our reponse to your comments 3,4 and 12.)
Full Screen / Esc
We will make the definitions clear on page 2763 line 10:
Printer-friendly Version
"Regional inorganic fertilizer consumption in 1995 is obtained from IFADATA (2011).
For the scenarios, we use a closed budget approach. For this purpose, we define
cropland soil Nr uptake efficiency (SNUE) as the ratio between Nr soil inputs (fertilizer,
manure, residues, atmospheric deposition, soil organic matter loss and free-living Nr
fixers) and soil withdrawals (harvest and crop residues minus seed and minus biological
C2464
Interactive Discussion
Discussion Paper
fixation by legumes and sugarcane). The definition of SNUE diverges from the widely
used Nr uptake efficiency (NUE) (Dawson et al., 2008) in the aspect that it regards only
Nr that is taken up from the soil. Seed Nr is already part of the plant biomass, and biological fixation of Nr from the atmosphere takes place within the organism. The reason
for the differentiation is that these Nr inputs should not subject to leaching, volatilisation
or denitrification prior to the uptake by the plant. This implies also, that legumes and
sugarcane have a higher NUE than other crops. SNUE is calculated on a regional level
for the year 1995 and becomes an exogenous scenario parameter for future estimates.
Its future development is determined by the scenario storyline (see chapter 2.4.).
In future scenarios, the soil withdrawals and the fixed SNUE determine the requirements for soil Nr inputs. If the amount of organic fertilizers is not sufficient, the model
can apply as much nitrogen fertilizer as it requires to balance out the budget. In our
model, the Nr inputs to crops have no influence on the yield. We assume in reverse
that a given crop yield can only be reached with sufficient Nr inputs. An eventual Nr
limitation is already reflected in the height of the crop-yield."
We will also discuss our definition in the new section 4.2 (see end of this manuscript
for a full version of the new chapter 4.2):
"Our closed budget approach to calculate future inorganic fertilizer consumption is
based on the concept of Soil Nr Uptake Efficiency (SNUE). Compared to other indicators of Nr efficiency that relate Nr inputs to crop biomass like Nr use efficiency (grain
dry matter divided by Nr inputs, not to be confused with Nr Uptake Efficiency), SNUE
has the advantage of an upper physical limit, as Nr withdrawals cannot exceed Nr inputs. At the same time, it indicates the fraction of losses connected to the application
of Nr inputs. As it includes a large number of Nr inputs, substitution effects can be
represented well. Finally, compared to Nitrogen Uptake Efficiency (NUE), one regional
value of SNUE suffices to simulate higher NUE of Nr fixing crops compared to normal
crops."
We will also replace the words Nr efficiency by the appropriate term (NUE or SNUE)
C2465
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
throughout the text.
In the following, we will answer to your comments 12 and 13, as they are on the same
topic.
12. On page 2763 the soil nitrogen use efficiency is defined as the ratio between Nr soil inputs (fertilizer, manure, residues, atmospheric deposition, soil
organic matter loss and free-living Nr fixers) and soil withdrawals (harvest and
crop residues minus seed and biological fixation by legumes and sugarcane). It
is not clear if biological N fixation by legumes and sugarcane is part of the withdrawal or not. I would argue that it is added to both the withdrawal AND to the
inputs.
BGD
9, C2463–C2503, 2012
Interactive
Comment
For the purpose of the model – to estimate the requirement of inorganic fertilizer and
the amount of Nr lost on the field – we think that biological fixation that takes place
within in the plant should be counter neither to soil inputs, nor to withdrawals.
We agree, that our definition of Soil Nr Uptake Efficiency (SNUE) diverges from the
commonly used definition of Nr Uptake Efficiency (NUE). We believe, that our definition
is better suited to estimate losses and can simulate the future dynamics of losses more
correctly.
The main reasoning behind our new definition is, that Nr fixed from the atmosphere
is not subject to losses prior to the uptake by the plant. While e.g. fertilizer leaches,
volatilizes or denitrifies when it is applied to fields, this is not the case for biologically
fixed Nr. Also Eggleston et al. (2006) assume, that Nr fixation itself has no significant
direct impact on losses and thus emissions, but only via the decay of residues.
The decay of residues contributes to losses also in our model. Straw and roots (from
both legumes and non-legumes) which remain on the field after harvest enter the soil
inputs N (xt )inp
(t,i) . All soil inputs exceeding the withdrawals are considered as losses
(eq. 28). Therefore, also Nr fixed biologically in straw and roots contributes to losses
when it decays on fields.
C2466
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
If biological fixation within the plant is not subject to losses, this also has implications
on the model dynamics. Assuming a SNUE of 50%, replacing one ton of biologically
fixed Nr requires two tons of other Nr inputs. Legumes and sugarcane therfore have
a higher NUE compared to normal crops. This was also observed in reality, see for
example Smil (1999).
9, C2463–C2503, 2012
We made this point more clear in the text on page 2763, as we already described in
the answer to question 2.
Interactive
Comment
BGD
13. The way the efficiency is calculated implies that any deficit is compensated
for by fertilizer nitrogen. In other words, deficits are assumed not to exist?
Deficits may well exist in our model.
The crop yields in MagPIE reflect the actual yield. If a yield is low, this may have its
origin in several climatic or management shortcomings, including the lack of nitrogen
fertilizers. It may therefore well be, that the crop yield is so low, because the plant is
Nr-deficient.
However, it would not be possible to reach these (even deficient) yields without sufficient Nr supply for these low yields. Nr requirements thus reflect Nr deficits, but Nr
deficits do not explicetly determine crop yields.
We made this point clear in the text on page 2763
"In future scenarios, the soil withdrawals and the fixed SNUE determine the requirements for soil Nr inputs. If the amount of organic fertilizers is not sufficient, the model
can apply as much nitrogen fertilizer as it requires to balance out the budget. In our
model, the Nr inputs to crops have no influence on the yield. We assume in reverse
that a given crop yield can only be reached with sufficient Nr inputs. An eventual Nr
limitation is already reflected in the height of the crop-yield.“
3. On page 2787 it is assumed that “no losses from the internally fixed Nr occurs, while the Nr fixed by free-living bacteria or in symbiosis with algae in rice
C2467
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
paddies is assumed to underly (?) the same proportion of losses as the other Nr
inputs”. It is absolutely unclear what is meant here. Is N fixation by free-living
bacteria lost? And is the N fixed by legumes not part of the soil-plant system,
where the straw and roots are decomposed after harvest and subject to losses?
But losses like surface runoff, denitrification and leaching are not computed directly.
Chapter A.3.4 is indeed confusing. We rewrite this chapter, concentrating on the technical implemention:
N ef f
"We calculate regional soil nitrogen uptake efficiency (SNUE) r(t,i)
by dividing total
inp
soil withdrawals N (xt )withd
(t,i) by total soil inputs N (xt )(t,i) .
eq A25
The soil inputs include inorganic fertilizer, manure, Nr released from soil organic matter loss, recycled crop residues, atmospheric deposition and Nr fixation by free-living
bacteria and algae. Nr in seed as well as Nr fixation by legumes and sugarcane are
not counted as soil inputs, as they reach the plant not via the soil. Soil withdrawals are
calculated by subtracting from the Nr in total plant biomass (harvested organ, aboveand belowground biomass) the amount of Nr that is not taken up from the soil and
therefore not subject to losses prior to uptake. The latter includes again seed Nr as
well as the Nr fixed from the atmosphere by legumes and sugarcane.
Eq A26 + A27
The loss of Nr from cropland soils N (xt )loss
(t,i) is defined as the surpluss of soil inputs
over soil withdrawals.
eq A28
As Nr from seed and biological fixation by legumes and sugarcane is integrated directly
into plant biomass, we assume that these inputs do not contribute to losses.
For the year 1995, we use historical data on regional fertilizer consumption based on
N ef f
N ef f
IFADATA (2011) to estimate r(t,i)
. In the following timesteps, r(t,i)
is fixed on an
ert
exogenous level (see Sect. A4), while fertilizer consumption N (xt )f(t,i)
becomes en-
C2468
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
dogenously calculated by the model."
4. In figure 2, N fixation is 25 Tg, of which 15 in belowground tissue. So I wonder
why this is not considered as being subject to losses, just like all other input
terms. Or is this amount from free-living bacteria and the fixation in rice paddies?
The fixation of 15 Tg Nr does not refer to free-living bacteria or fixation in rice paddies.
It is the Nr fixed in leguminous plants and sugarcane. To make this point more clear,
we changed the legend of figure 2 and seperate between "Biol. fixation by crops" and
"Other biol. fixation".
The N fixed within the plant is not subject to losses prior to decay, as we explained in
our answer to comment 3+12.
BGD
9, C2463–C2503, 2012
Interactive
Comment
5. The difference between the N fixation in the harvested parts in this study
and other recent papers needs more attention (different base year, plant growth
functions and N content of the grains), It is not clear what the reason is for this
large difference. How can plant growth functions be of influence, because it
is the production and N content that determine the result, and production is
(hopefully) equal to the data from statistics. If the N content is different, why is
that? And why not take the N content of Herridge in order to be consistent with
that inventory?
Our estimate for legume fixation (7 Tg) is similar to the estimate of Sheldrick et al.
(1996) (8 Tg) or Smil (1999) (10 Tg). However, it largely diverges from Herridge et al.
(2008), who estimated fixation by legumes to contribute 21 Tg Nr (not including fodder
or sugarcane).
There are several reasons, why we had lower total biological Nr fixation than Herridge
et al (2008). Their estimate is based on the following formula
Fix = prod · hi · %Nr · crop:shoot · %ndfa
C2469
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
with:
• Fix: Nr fixation
• prod: production
• hi: harvest index as ratio of shoot biomass to crop biomass
BGD
9, C2463–C2503, 2012
Interactive
Comment
• %Nr: Nr content of shoot
• crop:shoot: Ratio of total Nr (shoot + root) to shoot Nr.
• %ndfa: Percent of crop Nr derived by fixation from the atmosphere.
Because we think that the parameterization of Herridge overestimates Nr fixation, we
used an own parameterization for all of these factors, which mostly lead to lower estimates. We will explain this in the following on the example of soybean. Soybean
makes up more than 75% (16.44 Tg Nr) of total crop legume fixation in the estimate of
Herridge et al (2008), and is also responsible for the large difference in estimates.
• prod
There are two differences concerning the production.
Firstly, Herridge et al (2008) use a different base year (2005 instead of 1995).
During this time period, global soybean production increased by 69%. Shifting
the baseyear to 1995 reduces the estimate of Herridge et al by roughly 40%.
Secondly, Herridge et al. (2008) did not correct for dry matter when applying
their harvest index (they define it as dry matter grain to dry matter shoot ratio,
but seem to apply it to wet matter grain production if I compare their number to
FAOSTAT numbers). Correcting this error reduces their estimate of Nr fixation by
another 10%.
C2470
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
• hi We use different aboveground harvest indices to estimate total plant biomass.
Calculating our harvest index in 1995 in the same way as they do (dm grain to dm
shoot ratio), our global harvest index is 0.36 as opposed to 0.4 by Herridge et al.
(2008). The difference in harvest index actually decreases the difference in the
estimates from Herridge et al (2008) and our estimate. With our harvest index
the fixation rates are higher, as our shoot biomass is 2.75 times the grain mass,
while it is only 2.5 times in the case of Herridge et al (2008). This increases our
Nr fixation by 10% compared to their estimate.
BGD
9, C2463–C2503, 2012
Interactive
Comment
• %Nr We use a different Nr content of the shoot: 2.4% as opposed to 3% by
Herridge (2008). Our estimate is based on a relatively high value for soybean
grains (5.12%) based on Fritsch (2007) and a value of only 0.8% for soybean
residues (both Wirsenius 2000 and Eggleston 2006). This again results in 20%
lower Nr fixation compared to Herridge et al (2008).
• crop:shoot For belowground biomass our assumptions are strikingly different.
Again, we have a look at the prime fixation legume, soybean. We assume a DM
root to shoot ratio of 0.19:1 and a Nr-content of 0.8%, both in accordance with
Eggleston et al. (2006). Based on these assumptions, the global N root-to-shoot
ratio in 1995 is in our case 6.4%. Herridge et al assume it to be 50%. Unfortunately, Herridge et al. (2008) provides no citation or argument for his assumption
on the N root-to-shoot ratio. Stephen A. Williams, who carried out the literature
review for Eggleston et al. (2006), provides several citations for his estimate of
DM root to shoot ratio. I also tried to find some support for this ratio, finding for
example Sivakumar (1977), whose field observations showed with 0.11:1 a even
lower rootshoot ratio. Similarily, Dogan (2011) observed in field experiments ratios of 0.1:1 to 0.15:1. If we assume that the 0.19:1 DM root to shoot ratio is
correct, the belowground biomass of Herridge would require to have an N content
of 8%. This value is definetly too high. We used Egglestons (2007) estimate of
0.8% Nr per ton dry matter. This estimate is only based on a citation from the
C2471
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
year 1925. However, Dogan et al (2011) come in a recent study to similar results,
indicating N contents of 0.6 to 0.7%. We therefore believe that belowground N
estimates from Herridge et al are significantly too high, and remain with Egglestons estimates for root:shoot ratio and N contents of belowground biomass. The
differences in assumptions concerning the belowground biomass have a large
effect and reduce Nr fixation of soybeans by 30%. .
• %ndfa Sixthly, used the rates of Nr derived from fixation from the atmosphere
(%ndfa) estimates from farm observations (Peoples et al 2008 quoted in Herridge), while Herridget et al (2008) used values from experimental sites. The
farm observations were actually not based on Herridge et al. (2008), but only
quoted by Herridge and taken from Peoples et al (2008). For soybean, the main
fixer, these values were 0.58 instead of 0.68. This results in a 15% lower fixation.
BGD
9, C2463–C2503, 2012
Interactive
Comment
Aggregating the individual correction factors results in a reduction of Nr fixation from
soybeans from 16.44 Tg Nr to 4.6 Tg Nr.
16.44 Tg Nr · 0.6 · 0.9 · 1.1 · 0.8 · 0.7 · 0.85 = 4.6 Tg Nr
This explains the large divergence of our estimates.
The reviewer proposed we should consider to use the N-content of Herridge also
for our estimates. Unfortunately, this is not possible as Herridge et al (2006) only
considers the Nr content of the total AG shoot biomass, while we need a seperation
into harvested organ and crop residues.
However, for consistency reasons, we now changed our parameters to the values from
experimental sights that were also taken by Herridge et al. (2008). This also allow to
distinguish them regionally (see answer to question 6)
We now come to fixation by soybeans of 5.4 Tg Nr and a total Nr fixation by legumes
and sugarcane of 9.3 Tg Nr.
C2472
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
To point the differences in estimates more out, we change the discussion on page 2769
as follows:
"To estimate the fixation by legumes and sugarcane, we use a new approach based
on percentages of plant Nr derived from fixation, similar to Herridge et al (2008). This,
in combination with total above-and belowground Nr content of a plant, can predict
Nr fixation more accurately. However, we think that the parametrisation of Herridge
et al (2008) overestimates Nr fixation, especially for soybeans. Most importantly the
Nr content of the belowground residues as well as the shoot:root ratio seem too high
when comparing them with Eggleston et al (2006), Sivakumar et al (1977) or Dogan
et al (2011). But also the Nr content of the shoot seems too high given that soybean
residues have a much lower Nr content than the beans (Fritsch 2007, Wirsenius 2000
and Eggleston 2006). Correcting the estimates of Herridge et al (2008) for the water content further reduces their estimate. If one finally accounts for the difference in
baseyear between the two estimates, with global soybean production increasing by
69% between 1995 and 2005, we come to a global total fixation from legume and sugarcane of 9 Tg Nr in 1995 as opposed to 21 Tg Nr in 2005 in the case of Herridge et al.
(2008). Our estimate is inbetween the estimates of Smil (1999b) and Sheldrick (1996),
even though using a different approach."
BGD
9, C2463–C2503, 2012
Interactive
Comment
6. The N fixed by legumes is thus assumed to be in the harvested parts only.
This is not correct, since the N fixed is also in the straw and root tissue.
We do not assume that fixed Nr is only in the harvested parts.
We state this in several parts of the manuscript:
• In the methodology (p. 2762)
"The Nr fixed by leguminous crops and sugar cane is estimated by multiplying Nr
in total plant biomass (harvested organ, AG and BG residue) with plant specific
percentages of plant Nr derived from N2 fixation (Herridge et al., 2008)."
C2473
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
• in the discussion p2769
old version: "Our methodology uses the percentage of plant Nr derived from fixation. This, in combination with total above- and belowground Nr content of a
plant, can predict Nr fixation more accurately."
new version: "To estimate the fixatio by legumes and sugarcane, we use a new
approach based on percentages of plant Nr derived from fixation, similar to Herridge et al (2008). This, in combination with total above-and belowground Nr
content of a plant, can predict Nr fixation more accurately."
BGD
9, C2463–C2503, 2012
Interactive
Comment
• in the Appendix (p. 2785, 2786, 2787).
We do not know to which part of the text the Reviewer is referring. If there is a misleading formulation in the text, please indicate the page and line number so we can correct
this error.
Another problem is the nitrogen fertilizer applied to leguminous crops, which
may be considerable in some countries like Egypt. This means that the fractions
of nitrogen in the harvested parts which is assumed to be from nitrogen fixation,
is actually from fertilizer, so nitrogen fixation needs to be corrected for this.
We agree, that different management can lead to different rates of Nr fixation. Most importantly, if root nodules are not inocculated properly, or if too much inorganic fertilizer
is applied, Nr fixation rates can be reduced significantly. To account for the regional
variations in management, we change the model accordingly.
Management practices may vary considerably between regions. To account for this,
we use again the approach of Herridge et al (2008). The differentiation of soybean as
the largest single legume fixer is again of the largest importance. Of global soybean
production, four countries make up more than 85% of the producion: United States,
Brazil, Argentina and China (FAOSTAT 2012). We assume, based on Herridge, that
the percentage of crop Nr derived from biological fixation from the atmosphere (%ndfa)
in North America is 60%, in Latin America 80%, in Centrally planned Asia 50%. In
C2474
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
all other regions we assume the world average of 68%. The same values were also
applied to groundnuts, which have according to Herridge the same fixation rates. Similarily, we now distinguish fixation rates of sugarcane to 20% ndfa in Latin America and
10% ndfa in the rest of the world.
We make this change in methodology clear in the following parts of the paper: p.2762
line 26:
"The Nr fixed by leguminous crops and sugar cane is estimated by multiplying Nr in
total plant biomass (harvested organ, AG and BG residue) with regional plant specific
percentages of plant Nr derived from N2 fixation (Herridge et al., 2008)."
BGD
9, C2463–C2503, 2012
Interactive
Comment
p. 2785
"For legumes and sugar cane, where Nr fixation is the direct product of a symbiosis
of the microorganisms with the crop, we assumed that fixation rates are proportional
to the Nr in the plant biomass. The percentage of fixation-derived Nr is taken from
Herridge et al (2008). In the case of soybeans, groundnuts and sugarcane, fixation
rates vary between regions to account for differences in management practices like
fertilization or inocculation."
Finally, we change the table A.6 to include the regional values.
7.-9.
Comments 7-9 all refer to soil organic matter loss. We will answer each question
individually. However, since the whole description of the methodology in the appendix
on page 2786 line 10-25 changed, we will present the whole revised version of the text
in the end of question 9.
Also, a new table A7 (see attachment) will be included into the appendix, giving details
on the calculation of soil organic matter loss.
7. A further question is about the contribution from soil organic matter loss.
The authors state that they considered the conversion of forest and grassland
to cropland for the period 1980-1990. It is peculiar that 1980-1990 data is used
C2475
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
for 1995, since the cropland expansion for 1990-2000 is much less than in the
decade before.
According to Eggleston et al. (2006), the time of soil organic matter loss after conversion is approximately 20 years. Ideally, the period to estimate Nr release from soil
organic matter loss in 1995 would be 1975-1995.
If we use the HYDE data for 1990-2000 for the calculation of soil organic matter loss,
we come to very high numbers for soil organic matter loss (see new table below).
This is very surprising, because (as the reviewer mentioned), the land-expansion from
1990-2000 was rather low. The HYDE database in contrast has much higher rates of
land-expansion in this period. When we noticed this we asked Mr Kees Klein Goldewijk
for the reason. He answered us by email that FAO made an update of its database in
2011 and reduced its cropland area dramatically compared to the FAOSTAT estimate
in 2008. A corrected version of the HYDE database is not yet available.
We therefore decided not to use the period from 1990-2000.
We make this point clear in the rewritten part of the appendix, page 2786
"The results for the historical estimates can be found in Table A7. The estimates for
1990-2000 are too high. The HYDE estimates are based on an older release of FAOSTAT data, while more recent FAOSTAT data corrected the land-expansion significantly
downwards, reaching even a negative net-expansion for the period 1990-2000 (KleinGoldewijk, personal communication). For our estimate of Nr released by soil organic
matter, we used the estimates for the period 1980-1990."
In addition, FAO does not provide the conversion of grassland to cropland, so I
wonder where this information is coming from.
In our methodology, it is not important whether pasture or natural vegetation is
converted to cropland. According to Eggleston et al. (2006), soil carbon in pastureland
does not significantly differ from soil carbon in natural vegetation. Only after conversion
to cropland, a significant reduction takes place. It does therefore not matter whether
C2476
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
the cropland expansion goes into pasture or natural vegetation.
Our formulation in the Appendix, p.2786 is misleading. We change the paragraph
as follows: "When pastureland or natural vegetation is transformed to cropland, soil
organic matter is lost. This also releases Nr for agricultural production. As pastureland
and natural vegetation have a similar level of soil organic matter (Eggleston 2006), we
som ) on the basis of cropland
can calculate the Nr inputs from soil organic matter loss (N(t,i)
expansion, independent of whether this expansion occurs into natural vegetation or
pastureland.“
BGD
9, C2463–C2503, 2012
Interactive
Comment
The expansion of cropland according to HYDE is 69 million hectare for 19801990.
We used the HYDE database instead of FAOSTAT because this allows us to estimate
not only the net expansion of cropland, but the total land-expansion. If one hectar
is deforested in one cell and abandoned in a different cell of the same country,
the net-land expansion (thus the one captured by FAOSTAT) is zero. The brutto
land-expansion, which is important for the release of soil organic matter, however is 1
ha. This process is rather important: In the HYDE database, the net-expansion in the
period 1980-1990 is 69 ha, while the brutto land-expansion is 103 ha. As described in
our article (p.2786), we only counted the brutto-expansion that went into land that was
historically not used as cropland.
We change the text on p.2786 as follows
"As MAgPIE is calibrated to the cropland area in 1995, no land expansion occurs in this
timestep. Instead, we use the HYDE database with a 5’ resolution (Klein Goldewijk et
al., 2011) to estimate historical land expansion per grid-cell and the initial contribution
of soil organic matter loss to the cropland budget. The cropland area as covered by
FAOSTAT (2011) only allows to estimate the net-change of cropland area; expansion
C2477
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
and contraction of cropland area within the same country cancel out. The spatial explicit HYDE database allows to estimate actual land conversion, defined as the sum of
(positive) cropland expansion in each geographic grid-cell into land which was not used
as cropland since the year 1900. In the case that cropland area first shrinks and then
increases again, it is assumed that the same cropland area is taken into management
that was abandoned before, so that no new soil organic matter loss takes place. The
results for the historical estimates can be found in Table A7.“
BGD
9, C2463–C2503, 2012
Interactive
Comment
To make this point more clear, we add an extra column in the new table A7 (see attachment) indicating both net expansion and land conversion.
With a soil N loss of 28 Tg in 1995,
We made a mistake when we copied the value from our calculations into the paper. The
actual values was 25 Tg Nr (we used the correct value in the model for the projections).
We correct this value in our paper.
this means that 400 kg of N is lost per year and per hectare, or 4 tonnes of
nitrogen for a 10-year period. I assume that the soil organic matter loss after
the conversion to cropland takes 10 years. The soil organic carbon loss would
thus be 6100 kg C, or 12 tonnes of soil organic matter, or 120 tonnes of soil
organic matter over a 10-year period. As a global average this sounds like a
large amount, and this important term in the global cropland nitrogen balance
needs more explanation for readers to understand.
We agree that the amount of Nr released by soil organic matter loss is enormous. However the numbers of Nr release per ha and year are lower.
Firstly, the brutto area expanded is – as we noticed above – higher and reaches
103 Mha in the period 1980-1990. Therefore, the average amount released by landexpansion in the 10-year period 1980-1990 is 36 t C per ha or 2.4 t Nr per ha.
Secondly, this amount of Nr is not instantly released , but over a period of approximately 20 years (Eggleston 2006 Chapter 2 page 2.30). The annual release per ha
C2478
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
land-expansion is therefore in average 1.8 t C or 122 kg Nr per ha per year.
These calculations will be published in a new table A7 in the appendix. You find the
table attached to this document.
8. The inclusion of this soil loss term in the global nitrogen balance is fine, but it
needs to be stressed that this term does not add to the nitrogen to all croplands,
but only in those areas where recently forest has been cleared for cropland expansion. And since it is such a large amount that cannot be taken up by the
crops, most of it is probably lost by surface runoff, leaching and denitrification
and does not contribute to crop uptake. This will occur if this term is used in the
regional nitrogen budgets.
We agree with the reviewer, that the Nr release might surpass the amount of Nr that
can be taken up by crops. Also, it may distort the regional nitrogen efficiencies if the
amount of Nr that cannot be taken up by crops is significantly higher than in the case
of other Nr inputs like manure or fertilizers.
We therefore analyzed not only the regional average of Nr release, but also the variance
between grid-cells. If we limited the amount of Nr input per ha to 150 kg per ha and
year, the contribution of soil organic matter loss to the cropland budget was reduced
from 25.0 Tg to 21.9 Tg Nr per year. The reduction took mainly part region FSU, which
has very carbon-rich soils.
It is difficult to judge, whether the amount of losses is higher in the case of soil organic
matter loss than for other inputs. Inorganic fertilizer is also often applied in far too high
rates. In China, for example, average inorganic fertilizer application is 200 kg per ha,
with some regions exceeding 400 kg per ha (Smil 2002). Also, in regions with a strong
livestock sector , manure is applied in excessive rates because it cannot be transported
cheaply to where it is needed.
We therefore decided to keep the full amount of Nr from soil organic matter loss in our
budget, but to critically assess this point in our discussion:
C2479
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
p 2769 line 29
"...cropland has released 25 TgNr in 1995. With a yearly global average release of 122
kg Nr per ha newly converted cropland, the amount of Nr released will often exceed
the nutrients actually required by the crops, especially in temperate, carbon rich soils.
The cultivation of histosoils..."
9. For the future, this term is fixed exogenously according to the scenario assumptions. This sounds strange, because the model simulates land transformations, so can calculate forest conversion and thus the nitrogen from soil organic
matter loss, instead of the assumptions on page 2765. For example, the assumption that in the B1 and B2 scenarios, forest clearing will be halted may not
be correct if biofuel production will increase rapidly in future (which could be in
B1 and B2).
BGD
9, C2463–C2503, 2012
Interactive
Comment
We agree that an endogenous calculation of soil organic matter loss would be a large
model-improvement. We therefore take up the reviewer comment, and change the
model accordingly. Future soil organic matter loss is now determined by cropland
expansion in MAgPIE.
We describe our implementation in the manuscript and in the appendix:
We delete on P. 2765 the lines 17-19
"Similarly, we assume that conversion of natural vegetation and pasture into cropland,
leading to soil organic matter loss, will come to rest for the environmentally oriented
scenarios, whilst remaining constant in the economic oriented scenarios."
and insert instead on page 2763
"Nr release by the loss of soil organic matter after the conversion of pasture land or
natural vegetation to cropland is estimated based on the methodology of Eggleston et
al. (2006). Our estimates for 1995 use a dataset of soil carbon under natural vegetation
from the LPJmL model (Sitch et al., 2003; Gerten et al., 2004; Bondeau et al., 2007).
For 1995, we use historical land-expansion from the HYDE-database (Klein Goldewijk
C2480
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
et al., 2011), while the land-expansion in the future is estimated endogenously by MAgPIE."
In the appendix, we change the paragraph on page 2786 line 10 as follows:
"When pastureland or natural vegetation is transformed to cropland, soil organic matter
is lost. This also releases Nr for agricultural production. As pastureland and natural
vegetation have a similar level of soil organic matter (Eggleston 2006), we can calculate
som ) on the basis of cropland expansion,
the Nr inputs from soil organic matter loss (N(t,i)
independent of whether this expansion occurs into natural vegetation or pastureland.
Cropland expansion X(xt )areaexp
is calculated as the increase of X(xt )area
(t,j,v,w) into area
(t,j)
that has previously not been used as cropland. After the conversion of cropland, we
assume that cropland management releases 20–52% of the original soil carbon, depending on the climatic region (Eggleston et al., 2006), plus the full litter carbon stock
of the cell. Soil and litter carbon were estimated using the natural vegetation carbon
pools of LPJml. Yearly Nr losses after the potential conversion of an hectare cropland
som ) are then estimated on a cellular basis from the carbon losses, using a fixed C:N
(r(t,j)
ratio of 15 for the conversion of forest or grassland to cropland and dividing the result
by 20 years, which is the time assumed until the cabon stock arrives in the new equilibrium (Eggleston 2006). Soil organic matter loss is finally estimated by multiplying the
cropland expansion in each grid-cell with the yearly Nr losses per ha. For simplification,
we assume that instead of releasing the Nr in two consecutive timesteps (20 years),
the Nr is fully released in the timestep of conversion, and therefore multiply with 2.
som = X(x )areaexp · r som · 2
N(t,i)
t (t,j)
(t,j)
As MAgPIE is calibrated to the cropland area in 1995, no land expansion occurs in this
timestep. Instead, we use the HYDE database with a 5’ resolution (Klein Goldewijk et
al., 2011) to estimate historical land expansion per grid-cell and the initial contribution
of soil organic matter loss to the cropland budget. The cropland area as covered by
FAOSTAT (2011) only allows to estimate the net-change of cropland area; expansion
and contraction of cropland area within the same country cancel out. The spatial exC2481
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
plicit HYDE database allows to estimate actual land conversion, defined as the sum
of (positive) cropland expansion in each geographic grid-cell into land which was not
used as cropland since the year 1900. In the case that cropland area first shrinks and
then increases again, it is assumed that the same cropland area is taken into management that was abandoned before, so that no new soil organic matter loss takes place.
The results for the historical estimates can be found in Table A6. The estimates for
1990-2000 are too high. The HYDE estimates are based on an older release of FAOSTAT data, while more recent FAOSTAT data corrected the land-expansion significantly
downwards, reaching even a negative net-expansion for the period 1990-2000 (KleinGoldewijk, personal communication). For our estimate of Nr released by soil organic
matter, we used the estimates for the period 1980-1990."
BGD
9, C2463–C2503, 2012
Interactive
Comment
Estimates for future soil organic matter loss are included into the new parameter
overview table in the supplementary material.
Finally, we add a sentence to page 2771 line 9
“Nr release from soil organic matter (SOM) loss contributes to the Nr budget also in
the future, yet with lower rates. In the environmentally B oriented scenarios, cropland
expansion and therefore also SOM loss almost ceases due to forest protection, while
in the economically oriented scenarios, the loss of SOM still contributes 10 (A1) and
18 (A2) Tg Nr per year. In the A2 scenario the loss even continues at low rates until the
end of the century. The reduced inputs of soil organic matter loss have to be replaced
by inorganic fertilizers.”
10. Regarding the scale of the calculations, it is not clear if the nitrogen balances
are calculated on a regional basis, or spatially explicit. I assume that it is on a
regional basis, but this needs to be stated more clearly.
We agree with the reviewer that the scale of calculations for nitrogen balances has to
be more clearly described. We therefore will change the manuscript on p. 2763, line
10 (please note the changes made in response to your second comment) to:
C2482
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
“Regional inorganic fertilizer consumption in 1995 is obtained from IFADATA (2011). [...]
SNUE is calculated on a regional level for the year 1995 and becomes an exogenous
scenario parameter for future estimates.“
and page 2878 line 16
N ef f
"In the following timesteps, r(t,i)
is fixed on an exogenous level (see Sect. A4), while
the model can balance out the regional budget with inorganic fertiliser."
BGD
9, C2463–C2503, 2012
Interactive
Comment
11. If this is so, the nitrogen use efficiency is lumped at the scale of the 10
MagPie regions. This is another problem. If the crop mix changes, for example if
the share of nitrogen fixing crops increases, or the share of rice, the use of the
nitrogen use efficiency as a scenario variable is not correct.
It is one advantage of our definition of SNUE, that the different NUE of legumes and
non-legumes can be accounted for.
Assume the biomass of a lentil plant to contain 100 kg Nr. It receives approximately
50%, thus 50kg of its Nr from biological fixation. The remaining 50kg have to stem
from external sources. With a SNUE of 50% this requires fertilizers with 100kg Nr.
The NUE is thus 100/(50+100) = 75%, much higher than the average crop where it is
50%. These higher NUE of legumes and forages have also been observed in realitiy.
Smil (1999) estimates that normal crops have a NUE of 0.35-0.55, while legumes and
forages have uptake efficiencies of 0.65 and 0.75 respectively.
Nevertheless, we agree that many other processes that determine SNUE are not covered within the model. Different crop species have a different ability taking up available
soil Nr. Similarily, different types of Nr fertilizers (Manure, residues, inorganic fertilizer)
can be taken up better or worse. Thus a change in composition of crop species or Nr
inputs should have an influence on Nr efficiency. Besides, there are numerous other
factors influencing SNUE like management or climate.
Our approach, assuming that every Nr input will be taken up with the same Nr efficiency
is thus a simplification. However, the processes that determine Nr uptake efficiency are
C2483
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
very complex. This is one reason, why we made it an exogenous scenario parameter.
This makes our underlying assumptions clear.
BGD
We will discuss this in the new chapter 4.2. which critically assesses our scenario
assumptions (see end of this manuscript for a full version of the new chapter 4.2).
9, C2463–C2503, 2012
"The future development of SNUE is highly uncertain. It depends on numerous factors:
most importantly on the management practices like timing placing and dosing of fertilizers and the use of nutrient trap crops. Also a general improvement of agricultural
practices like providing adequate moisture and sufficient macro- and micronutrients,
pest control and avoiding soil erosion can contribute their parts. Finally, climate, soils,
crop varieties and the type of nutrient inputs also influence Nr uptake efficiency. The
complexity of these dynamics and the numerous drivers involved still do not allow to
make long-term model estimates for Nr efficiencies, but should be target for future research. Meanwhile, we use SNUE as an explicitly defined scenario parameter. As it
descriptively indicates the share of losses, and as the theoretical upper limit of 1 is
clearly fixed, it makes our model assumptions transparent and easily communicable.“
Interactive
Comment
Comment 12 and 13 were answered earlier (see above)
14. Although the Appendix presents many details about the methods, data and
model, it is not possible to understand how the model works and how scenarios
were constructed. For example, the food demand is expressed as EJ (figure
A2). Readers would be grateful to see the actual regional domestic production in
tonnes, and also the cropland areas, yield development and fertilizer use. That
would also clarify a bit more why fertilizer use increases so rapidly under A1
assumptions, and decreases so rapidly in B2 and not in B1.
In response to this valuable comment we now want to include a large list of model
outputs into the supplementary material, including domestic production in tons and Nr,
cropland areas, yield development and fertilizer use.
The table also includes the distribution of production to food, feed, material demand,
C2484
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
seed, exports. It also includes all cropland inputs and withdrawals, such that model
dynamics can be easily traced back.
BGD
Also, the new section of the discussion 4.2 will assess the implications of our scenario
assumptions (see end of this manuscript for a full version of the new chapter 4.2)
9, C2463–C2503, 2012
15. In the A1 scenario fertilizer use increases very rapidly. This is peculiar, because at present fertilizer use efficiency is increasing in industrialized countries,
and one could expect that the same will happen in developing countries. China
and India are currently the major consumers of fertilizer nitrogen, determining
a large part of the global total and its increase. When China and India develop,
subsidies will probably decrease or be stopped, so the efficiency will have to
increase. This will not be only in the environmentally oriented scenarios.
Interactive
Comment
The strong increase of fertilizer use in the A1 scenario has several reasons. Most importantly, the strong increase in income leads to higher total consumption, and also
to an increasing consumption of livestock products. The latter requires the production
of feedstock. Along with strong economic growth goes the rapid modernization of the
livestock sector with higher fraction of Nr rich concentrate feed and conversion byproducts. This leads to a very high total production.
The increasing soil Nr uptake efficiency, which is also part of the storyline, is slowing
down the increase in fertilizer consumption. As shown in table 1, our scenario assumptions include a strong increase in Nr uptake efficiency in all scenarios. In the B
scenarios it reaches levels that can be seen as very optimistic, as no current region
comes close to this value presently.
We also assume a catch-up concerning the efficiency, such that all regions reach the
same NUE in 2045. E.g. China wil therefore have much higher growth rates in efficiency than the Europen Union.
To make the development of nitrogen efficiency more clear, we will also include thes
values both globally and regionally into the new tables in the supplementary material.
C2485
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
Additional to the new table in the supplementary material, the new chapter 4.2 (see
end of this manuscript for a full version of the new chapter 4.2) explains the implications
of our scenario assumptions and discusses the future development of Soil Nr uptake
efficiency (SNUE).
16. Page 2771, last paragraph: global simulated fertilizer use for 2005 agrees
with the statistics, but the regional results do not. This is rather frightening,
if after 10 years the model already deviates from the data, because the effect of
wrong scenario assumptions will be dramatic after 100 years. The regions where
fertilizer use is underestimated are those where nitrogen use is most important
globally (see also 15 above), and this means a considerable overestimation in
other regions. It is also interesting to discuss the large difference between IFA
and FAO regional data, which seems to be especially large in the period 20052010.
We agree, that our regional estimates are of a lower quality than the global estimates.
These variations can be largely explained by trade dynamics. These dynamics are
extremely difficult to predict (and our model certainly has a simplified representation
compared to more economically oriented CGE models), but have a strong impact on
regional results. However, the global estimates are not necessarily influenced by this.
If grains are exported from Europe to North America, more fertilizer will be used in
Europe, but less in North America. These two trends cancel out more or less on a
global level, but have a strong effect on the regional budgets.
In the new section 4.2 (see end of this manuscript for a full version of the new chapter
4.2), we enter the following text:
“We assumed that trade liberalisation continues in all scenarios, even though at different paces. The trade patterns diverge strongly between the scenarios, even though
certain dynamics persist: Subsaharan Africa, Europe and Latin America tend to become lifestock exporting regions, while South, Central and South East Asia as well as
the Middle East and Northern Africa become importers of livestock products. On the
C2486
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
other hand, Subsaharan Africa and Pacific Asia become importers of crop products,
while the Former Soviet Union and Australia become exporter of crops. Trade dynamics in MAgPIE are determined partly on the basis of historical trade patterns, partly
by competitiveness. However, certain other dynamics that are of great importance in
reality, most importantly political decisions like tariffs or export subsidies are not represented explicitly in the model. Due to the uncertainty regarding trade patterns, regional
production estimates are therefore of higher uncertainty than global estimates. Trade
patterns have strong implications on the Nr cycle. As soon as two regions are trading, the fertilizer consumption also shifts from the importing to the exporting country.
Even more, Sub-Saharan Africa currently imports crops and exports livestock products.
Livestock fed with imported crops contributes in the form of manure to the cropland soil
budgets and facilitates Sub-Saharan Africa to use little inorganic fertilizer. Also in our
future scenarios, the African livestock sector is very competitive and the inorganic fertilizer consumption does not increase until the mid of the century. A similar dynamic
can be observed in Latin America, where inorganic fertilizer consumption also stays
rather low.“
Trade is not the only origin of uncertainty. Also our underlying GDP and population
scenarios are itself scenarios with high uncertainty. SRES GDP estimates in China
diverge by factor 0.4 to 1.1 from actual GDP measured in 2010, in MEA by factor 1.2 to
1.8. As our scenarios are just representative pathways, I think that they are not outside
the uncertainty range.
Moreover, after the diverse changes to the model undertaken for the review, our results
are now differnt.
Inorganic fertilizer consumption in Europe stagnates or only slightly increases in all
scenarios, which is absolutely in line with the development since the 90s. Also China
shows a continuous increase in consumption, even though at slower pace than in the
last decades.
In India, fertilizer consumption rose sharply in the last decades, while our projections
C2487
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
suggest a stagnating or only little increasing fertilizer consumption. Pacific OECD
(Japan, New Zealand and Australia) in contrast, arise as new exporter of crop products
and have a strongly increasing fertilizer demand.
We will rewrite the section (page 2771 last paragraph) as follows, also taking up your
comment on the differences between IFADATA and FAOSTAT (which was new to us):
“Data on historic fertilizer consumption is provided by IFADATA (2011) and FAOSTAT
(2012). Both estimates diverge, as they use different data sources and calendar years.
On a regional level, differences can be substantial. FAOs estimate for fertilizer consumption in China in the year 2002 is 13% higher than the estimate by IFA. As IFADATA (2011) provides longer continuous time-series, we will refer to this dataset in the
following. Fertilizer consumption between 1995 and 2009 (IFADATA, 2011) grows by
+1.8% per year. The estimates of Daberkow et al. (2000) and Bouwman et al (2009,
2011) are lower, with average growth rates of -0.4% to + 1.7% over the regarded period
of 20 to 50 years. Our 50 year average growth rate also stays with +0.9% to +1.7%
below the observations. Yet, our short-term growth rate from 1995 to 2005 captures
the observed development with a range of +1.5% to +2.4% between the scenarios.
Due to trade (see “[new]” section 4.2), our regional fertilizer projections are more uncertain than the global ones. However, our results still meet the actual consumption
trends of the last decades for most regions. However, fertilizer consumption in India
rises slower than in the past or even stagnates, while the Pacific OECD region shows
a strong increase in fertilizer consumption.”
Some minor comments are: - Page 2759, last para: the existing model must be
extended, or has been ?
We changed it to "has been".
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Atmospheric deposition: is it the same for all scenarios? That is strange, because the scenarios are so different, probably also causing differences in emissions and thus deposition.
C2488
Discussion Paper
We agree that atmospheric deposition should be consistent with Nr emissions in each
scenario.
Therefore we decided to connect the growth rate of atmospheric deposition to the
growth rate of the amount of Nr that volatilizes in the form of NOx and NHy during
cropland fertilization.
In reality, NOx and NHy emissions from industry and traffic, also determine the deposition rates. However, as a major part of volatilized Nr will be deposited close to the
emission source, the largest part of cropland atmospheric deposition should also stem
from agricultural Nox and Nhx emissions.
BGD
9, C2463–C2503, 2012
Interactive
Comment
P2763 line 5
"The regional amount of atmospheric deposition on croplands for 1995 is taken from
Dentener (2006). For future scenarios, we assume that the atmospheric deposition
per cropland area grows with the same growth rate as the average regional agricultural
NOx and NHy emissions (see chapter 2.3.4)."
P. 2785 line 4 ff
"A major part of the Nr lost from field in the form of NOx and NHy as well as other Nr
compounds from the combustion of fossil fuels are lateron deposited from the atmosphere on cropland area. Based on spatial datasets for atmospheric deposition rates
(Dentener, 2006) and cropland area (Klein Goldewijk et al., 2011), we derive average
atmospheric deposition rates per area for each region (ridep ). For the future we assume,
that these deposition rates grow with the same growth rate as the agricultural NOx and
NHy emissions N (xt )volat
(t,i) (see A35 in section A.3.5). This implies the assumption that
cropland deposition of NOx and NHy from non-agricultural sources grow with the same
growth rate as agricultural NOx and NHy emissions. As a large part of volatilized Nr
will be deposited close to the emission source, the largest part of cropland atmospheric
deposition probably stems from agricultural NOx and NHx.
P
dep
volat
volat
area
"
N (xt )dep
(j,v,w) X(xt )(t−1,j,v,w) · ri
(t,i) = N (xt )(t,i) )/(N (xt )(t−1,i) ·
C2489
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
finally on page 2788, we add the following lines after line 15:
"The NOx and NHy volatilisation on cropland area N (xt )volat
t,i , which is required for
the calculation of atmospheric deposition in formula A23, section A.3, is calculated as
follows
(gas_f ert)
f ert
ex
N (xt )volat
+ (N (xt )( t, i)m + N (xt )ex
(t,i) = N (xt )(t,i) · ri
(t,i,l,grazp) + N (xt )(t,i,l,grazc) ) ·
gas_m
gas
_
awms
r
+ (N (x )ex
·r
"
i
t (t,i,l,house)
i
BGD
9, C2463–C2503, 2012
Interactive
Comment
We come to very similar results than Dentener (2006), with a global total of 26-31 Tg
Nr in 2045 (old implementation 27-30 Tg Nr). Also on a regional level, atmospheric
deposition rates match very well, only in AFR deposition is higher and in SAS it is
lower compared to the old implementation.
Our estimates for atmospheric deposition can be found in the new additional output
table in the supplementary material.
- Page 2769, last para: N accumulation is not considered in this paper.
We agree that N accumulation is not considered. We should make more clear that our
calculations of soil organic matter loss only estimate soil Nr depletion, not accumulation.
We therefore change the sentence as follows
"Accumulation or depletion of Nr in soils has so far been neglected in future scenarios (Bouwman et al., 2009, 2011), assuming that soil organic matter reached a stable
equilibrium and all excessive Nr will volatilize or leach. However, the assumption of a
steady state for soil organic matter should not be valid for land conversion and cultivation of histosoils. Our rough bottom-up calculations estimate that the the depletion of
soil organic matter after transformation of natural vegetation or pasture to cropland has
released 25 Tg Nr in 1995. ..."
- Page 2771, first para: using constant excretion rates implies that the feed conversion ratio decreases and excretion per unit of product decreases, which reC2490
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
flects an improving efficiency. Simply assuming an increasing efficiency sounds
counterintuitive.
We do not agree that an increasing feeding efficiency is counterintuitive. In an interregional comparison one can well see that developed countries have higher feeding
efficiencies.
However we should make more clear, on which assumptions both methods are based,
and what are eventual shortcomings.
BGD
9, C2463–C2503, 2012
Interactive
Comment
We think that the implementation by Bouwman et al. (2011) underestimates manure
excretion. While the modernization of the livestock sector should indeed decrease the
amount of manure per livestock unit, this does not mean that productivity improvements are not connected to changing excretion rates at all. IPCC (1996) for example
estimates excretion rates of dairy cattle to be 60 kg Nr per animal in Africa, while in
North America, excretion rates are 100 kg Nr per animal. At the same time, the excretion per livestock product is 1.75 kg Nr / t DM in Africa and 0.4 kg Nr/ tDM in North
America. Productivity improvements are therefore connected both to an decrease in
excretion per animal products and an increase in excretion per animal.
In contrast to Bouwman et al (2011), our methodology has the disadvantage that while
all regions converge towards European productivity, no productivity improvements beyond the European level are possible. This in turn translates into an overestimation of
manure excretion. We make this point more clear in the new chapter 4.2
"Concerning the productivity of the livestock sector, we assume that the feed required
to produce one ton of livestock product is decreasing in all scenarios, even though at
different rates. Starting from a global level 0.62 kg N in feed per ton livestock product
dry matter, the ratio decreases to 0.4 (A1) or 0.52 (B2) in 2095 (see supplementary
material). A critical aspect is, that as all regions convert towards the European feed
baskets, no productivity improvements beyond the European level take place. Beside
the improvement of feed baskets , the amount of feed is also determined by the mix
C2491
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
of livestock products, with milk and eggs requiring less Nr in feed than meat. As we
could not find any noticable historical trend in the mix of products [FAOSTAT, 2011],
we assumed that current shares remain constant in the future. This causes continuing
high feeding efficiencies in Europe and North America, where the share of milk and
non-ruminant meat is high.“
and on p. 2771 line 4:
"Secondly, Bouwman et al. (2011) assume rising Nr excretion rates per animal for
the past, but constant rates for the future, such that weight gains of animals are not
connected to higher excretion rates. As the current excretion rates in developing regions are still lower than in developed regions (IPCC 1996), this assumption might
underestimate the growth of excretion rates in developing regions. Our implementation
calculates excretion rates based on the feed baskets and the Nr in livestock products. Under the assumption that developing countries increasingly adopt the feeding
practices of Europe, this top-down approach results in increasing excretion rates per
animal in developing regions. However, as we assume no productivity improvements
in developed countries, we tend to overestimate future manure excretion in developed
countries."
BGD
9, C2463–C2503, 2012
Interactive
Comment
- Page 2779: the IMAGE model also accounts for nutrient withdrawal by fodder
crops.
I assume that in the context of page 2779, you mean that the IMAGE model also
accounts for conversion byproducts. We were not aware of this, and therefore we
delete the sentence
"So far, they have not been accounted for in most global material flow analysis, an
exception being Wirsenius (2000) and Weindl et al. (2010)."
As we also write on page 2757 line 16 that
"Above all, most studies do not consider fodder crops and belowground residues as
major Nr withdrawals from cropland soils."
C2492
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
we will also change this sentence. Even though in earlier estimates like Sheldrick
(1996) or Liu (2010) fodder crops were not included, large models like the IMAGE
model probably account for fodder crops. We therefore reformulate the paragraph more
conservative:
"Belowground residues were so far not considered explicitly by other global studies,
even though they withdraw large amounts of Nr from soils, and their decay on fields
contributes to Nr losses and emissions. Similary, not all past studies included fodder
crops into their budgets, although they make up a considerable share of total cropland
production."
BGD
9, C2463–C2503, 2012
Interactive
Comment
- Page 2791, line 19-23: a comparison with 2012 food demand statistics is needed
here.
The comparison of actual food demand and simulated food demand is done in figures
A2 and A3. The past observations overlap with our simulated results in the year 19902009. In addition, we now add the following sentence:
"The food demand projections are based on population and income growth of the SRES
scenarios. As can be seen in figure A2 and A3, the historical data is met more or less
precisely depending on the scenario. Global food calory demand diverges in 2005
by 98 PJ (+0.4%) (B1) to 452 PJ (1.7%) (B1), while meat demand diverges by -244
PJ (-5.2%) (A2) to +60 PJ (1.2%) (B2). The largest differences can be observed in
the estimates for CPA, where the A2 scenario diverges by -422 PJ (-31.5%) while the
B2 scenario almost matches the observed data with 15 PJ (+1.1%). Large parts of
these variations in estimates are determined by the uncertainty of the original SRES
projections for population and GDP."
The newest available year for FAOSTAT food supply statistics is 2007. As the SRES
data is only available in 5 year timesteps, we compared the data to 2005.
- Table 1: fertilizer use in some world regions is insignificant. When fertilizer use
increases in such places, efficiency will go down.
C2493
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
We do not agree, that efficiency should necessarily go down when fertilizer use increases.
Certainly, the Nr efficiency of a farming system would go down if additional fertilizer
was applied with all other parameters staying equal.
But in our model, the increased fertilizer application is a consequence of increased production, which requires larger amounts of fertilizers. Thereby, the amount of fertilizer
has no influence on the crop yields, but the given crop yield and the given soil nutrient
uptake efficience determine the amount of Nr, that is required for fertilization.
BGD
9, C2463–C2503, 2012
Interactive
Comment
We will make this point more clear in the last paragraph of section 2.3.3 on page 2763 In
future scenarios, the soil withdrawals and the fixed SNUE determine the requirements
for soil Nr inputs. If the amount of organic fertilizers is not sufficient, the model can
apply as much nitrogen fertilizer as it requires to balance out the budget. In our model,
the Nr inputs to crops have no influence on the yield. We assume in reverse that a
given crop yield can only be reached with sufficient Nr inputs. An eventual Nr limitation
is already reflected in the height of the crop-yield.“
I do not know how the soil organic matter loss would add to the inputs and determine the efficiency, but this needs explanation on a regional level, for example
with a few examples.
Soil Nr uptake efficiency (SNUE) is deteminded in 1995 as described in the Appendix
A3.4. As this paragraph was not written very well, we rewrote it (see answer to comment 3). In table 3, which includes all soil inputs and withdrawals on a global scale, we
replace the name of the line "IN/OUT" and "IN*/OUT*" by "NUE" and "NUE*". Additionally, we add two rows with "SNUE" and "SNUE*".
Finally, in the new supplementary material, all Nr inputs and withdrawals which are
used for the calculation of NUE and SNUE in 1995 are listed, both globally and regionally. This table allows to see, how large the contribution of organic matter loss, fertilizer
or residues is to the regional budgets.
C2494
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
As we referred several times only to fragments of the new chapter 4.2, we will in the
following show the full chapter 4.2 in one piece. It shall be located on page 2770,
between the chapters “The current state of the agricultural Nr cycle” and “The future
expansion of the Nr cycle”.
"4.2. Critical assessment of scenario assumptions
The simulation of the widely used SRES storylines (Nakicenovic et al., 2000) facilitates the comparison with other studies like Bouwman et al. (2009) or Erisman et al.
(2008) and allows for the integration of our results into other assessments. However,
the SRES storylines provide only a qualitative description of the future. In the following,
the key assumptions underlying our parametrisation shall be discussed.
All SRES storylines tend to assume a continuation of current trends, without external shocks or abrupt changes of dynamics. They merely diverge in the interpretation
of past dynamics or the magnitude of change assigned to certain trends. Population
grows at least until the mid of the 21st century, and declines first in developed countries.
Per-capita income grows throughout the century in all scenarios and all world regions,
and developing regions tend to have higher growth rates than developed regions. This
has strong implication on the food demand, which is driven by both population and income growth. Due to the Engels-shaped demand curve for food, the development of
total food demand depends mostly on the income growth of low-income regions. The
same holds for the share of animal calories. In the first half of the century population
growth does not differ much in most regions. The pressure from food demand is therefore highest in the high-income A1 scenario, while in the second half, the A2 scenario
also reaches a medium income and therefore a relatively high per capita-demand. This,
combined with a high population growth leads to a very high total food demand in the
A2 scenario. As food demand is exogenous to our model, price effects on consumption
are not captured by the model. However, even in the A2 scenario the shadow prices
(Lagrange multipliers) of our demand constraints increase globaly by 0.5% per year
until 2045, with no region showing higher rates than 1.1%. This indicates only modest
price pressure, lagging far behind income growth.
C2495
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
Concerning the productivity of the livestock sector, we assume that the feed required
to produce one ton of livestock product is decreasing in all scenarios, even though at
different rates. Starting from a global level 0.62 kg N in feed per ton livestock product
dry matter, the ratio decreases to 0.4 (A1) or 0.52 (B2) in 2095 (see supplementary
material). A critical aspect is, that as all regions convert towards the European feed
baskets, no productivity improvements beyond the European level take place. Beside
the improvement of feed baskets, the amount of feed is also determined by the mix
of livestock products, with milk and eggs requiring less Nr in feed than meat. As we
could not find any noticable historical trend in the mix of products [FAOSTAT, 2011],
we assumed that current shares remain constant in the future. This causes continuing
high feeding efficiencies in Europe and North America, where the share of milk and
non-ruminant meat is high.
As we calculate our livestock excretion rates based on the feedmix, the increased feeding efficiency also translates into lower manure production per ton livestock product.
At the same time, our scenarios assumptions of an increasing share of either anaerobic digesters or daily spread in manure management also lead to higher recycling
rates of manure excreted in confinement. Even though with increasing development
an increasing share of collected manure is applied also to pastureland as opposed to
cropland, the amount of manure Nr applied to crops remains rather constant per t DM
crop biomass. Due to the increasing Nr efficiency, its ratio relative to other Nr inputs
like inorganic fertilizers increases.
Our closed budget approach to calculate future inorganic fertilizer consumption is
based on the concept of Soil Nr Uptake Efficiency (SNUE). Compared to other indicators of Nr efficiency that relate Nr inputs to crop biomass like Nr use efficiency (grain
dry matter divided by Nr inputs, not to be confused with Nr Uptake Efficiency), SNUE
has the advantage of an upper physical limit, as Nr withdrawals cannot exceed Nr inputs. At the same time, it indicates the fraction of losses connected to the application
of Nr inputs. As it includes a large number of Nr inputs, substitution effects can be
represented well. Finally, compared to Nitrogen Uptake Efficiency (NUE), one regional
C2496
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
value of SNUE suffices to simulate higher NUE of Nr fixing crops compared to normal
crops.
The level of SNUE is in our model an exogenous scenario parameter for future simulations, which has enormous impact on the estimates of inorganic fertilizer consumption
and N2O emissions. If SNUE would be 5 percentage points lower, fertilizer consumption would increase by 8 to 10% in 2045, depending on the scenario. At the same time,
total agricultural N2O emissions would increase by 11 to 15%. If fertilizer efficiency
would increase by 5 percentage points, fertilizer consumption would fall by 7 to 8%
and emissions would decrease by 9 to 13%. As the magnitude of Nr flows is higher in
some scenarios, a +-5% variation of SNUE translates in the A1 scenario into a change
of fertilizer consumption of -32 to +37 Tg Nr and a change of -1.06 to +1.26 Tg N2O-N
of emissions in 2045, while in the B2 scenario fertilizer changes only by -20 to +24 Tg
Nr and emissions by -0.7 to +0.8 Tg N2O-N.
The future development of SNUE is highly uncertain. It depends on numerous factors:
most importantly on the management practices like timing placing and dosing of fertilizers and the use of nutrient trap crops. Also a general improvement of agricultural
practices like providing adequate moisture and sufficient macro- and micronutrients,
pest control and avoiding soil erosion can contribute their parts. Finally, climate, soils,
crop varieties and the type of nutrient inputs also influence Nr uptake efficiency. The
complexity of these dynamics and the numerous drivers involved still do not allow to
make long-term model estimates for Nr efficiencies, but should be target for future research. Meanwhile, we use SNUE as an explicitly defined scenario parameter. As it
descriptively indicates the share of losses, and as the theoretical upper limit of 1 is
clearly fixed, it makes our model assumptions transparent and easily communicable.
Our assumptions concerning the development of SNUE are rather optimistic. In 1995,
none of the 10 world regions reached a SNUE of 60%, and four regions (CPU, FSU,
PAS, SAS) were even below 50%. The current difference between the region with the
lowest SNUE (CPA with 43%) and the region with the highest SNUE (EUR with 57%)
is thereby still lower than the difference of EUR and our scenario parameter of 70% for
C2497
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
the environmentally oriented scenarios.
We assumed that trade liberalisation continues in all scenarios, even though at different paces. The trade patterns diverge strongly between the scenarios, even though
certain dynamics persist: Subsaharan Africa, Europe and Latin America tend to become lifestock exporting regions, while South, Central and South East Asia as well as
the Middle East and Northern Africa become importers of livestock products. On the
other hand, Subsaharan Africa and Pacific Asia become importers of crop products,
while the Former Soviet Union and Australia become exporter of crops. Trade dynamics in MAgPIE are determined partly on the basis of historical trade patterns, partly
by competitiveness. However, certain other dynamics that are of great importance in
reality, most importantly political decisions like tariffs or export subsidies are not represented explicitly in the model. Due to the uncertainty regarding trade patterns, regional
production estimates are therefore of higher uncertainty than global estimates. Trade
patterns have strong implications on the Nr cycle. As soon as two regions are trading, the fertilizer consumption also shifts from the importing to the exporting country.
Even more, Sub-Saharan Africa currently imports crops and exports livestock products.
Livestock fed with imported crops contributes in the form of manure to the cropland soil
budgets and facilitates Sub-Saharan Africa to use little inorganic fertilizer. Also in our
future scenarios, the African livestock sector is very competitive and the inorganic fertilizer consumption does not increase until the mid of the century. A similar dynamic
can be observed in Latin America, where inorganic fertilizer consumption also stays
rather low.
In the environementally oriented scenarios B1 and B2, we restricted the expansion of
cropland into intact and frontier forest. However, these protected areas only include
some of the most vulnerable forest areas, and its implementation is assumed to take
place gradually until 2045. Large forest areas are still cleared, most importantly in the
Congo river basin and the southern part of the Amazonian rainforest. Due to the land
restrictions, crop yields have to increase faster to be able to settle the demand with the
available cropland area. Also, the area-dependent Nr inputs from soil organic matter
C2498
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
loss, atmospheric deposition and free-living bacteria are lower. "
Bibliography
Bondeau, A., P. C Smith, S. Zaehle, S. Schaphoff, W. Lucht, W. Cramer, D. Gerten,
et al. 2007. “Modelling the Role of Agriculture for the 20th Century Global Terrestrial
Carbon Balance.” Global Change Biology 13: 679–706.
Bouwman, A. F, A.H.W. Beusen, and G. Billen. 2009. “Human Alteration of the Global
Nitrogen and Phosphorus Soil Balances for the Period 1970–2050.” Global Biogeochemical Cycles 23: 1–15.
BGD
9, C2463–C2503, 2012
Interactive
Comment
Bouwman, L., K. K. Goldewijk, K. W. Van Der Hoek, A. H. W. Beusen, D. P. Van Vuuren,
J.
Willems, M. C. Rufino, and E. Stehfest. 2011. “Exploring Global Changes in Nitrogen
and Phosphorus Cycles in Agriculture Induced by Livestock Production over the 1900–
2050 Period.” Livestock and Global Change Special Feature of Proceedings of the
National Academy of Sciences: 1–6. doi:10.1073/pnas.1012878108.
Daberkow, S., J. Poulisse, and H. Vroomen. 2000. Fertilizer Requirements in 2015
and 2030. Rome: Food and Agriculture Organization of the United Nations(FAO).
Dawson, Julie C., David R. Huggins, and Stephen S. Jones. 2008. “Characterizing
Nitrogen Use Efficiency in Natural and Agricultural Ecosystems to Improve the Performance of Cereal Crops in Low-input and Organic Agricultural Systems.” Field Crops
Research 107 (2) (May 10): 89–101. doi:10.1016/j.fcr.2008.01.001.
Dentener, F. 2006. “Global Maps of Atmpshperic Nitrogen Deposition, 1860, 1993
and 2050.” Oak Ridge, Tennessee, U.S.A.: Oak Ridge National Laboratory Distributed
Active Archive Center. http://daac.ornl.gov/.
Dogan, K., I. Celik, M. Gok, and A. Coskan. 2011. “Effect of Different Soil Tillage Methods on Rhizobial Nodulation, Biyomas and Nitrogen Content of Second Crop Soybean.”
C2499
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
African Journal of Microbiology Research 5 (20): 3186–3194.
Eggleston, H. S., L. Buendia, K. Miwa, T. Ngara, K. Tanabe, and K. Hayama, eds. 2006.
2006 Guidelines for National Greenhouse Gas Inventories, Prepared by the National
Greenhouse Gas Inventories Programme. Kanagawa, Japan: Institute for Global Environmental Strategies.
FAOSTAT. 2011. “Database Collection of the Food and Agriculture Organization of the
United Nations.” www.faostat.fao.org.
BGD
9, C2463–C2503, 2012
Interactive
Comment
Fritsch, Friedhelm. 2007. Nährstoffgehalte in Düngemitteln Und Im Erntegut; Für
Die Düngeplanung; Für Nährstoffvergleiche. Bad Kreuznach: Dienstleistungszentrum
Ländlicher Raum Rheinhessen-Nahe-Hunsrück.
Gerten, Dieter, S. Schaphoff, U. Haberlandt, Wolfgang Lucht, and S. Sitch. 2004. “Terrestrial Vegetation and Water Balance. Hydrological Evaluation of a Dynamic Global
Vegetation Model.” Journal of Hydrology 286: 249–270.
Herridge, David F., Mark B. Peoples, and Robert M. Boddey. 2008. “Global Inputs of
Biological Nitrogen Fixation in Agricultural Systems.” Plant and Soil 311 (1-2) (July):
1–18. doi:10.1007/s11104-008-9668-3.
IFADATA. 2011. “Statistical Database of the International Fertilizer Association (IFA).”
www.fertilizer.org/ifa/ifadata/.
IPCC. 1996. “Volume 2: Workboook, Chapter 4: Agriculture.” In Revised 1996 IPCC
Guidelines for National Greenhouse Gas Inventories. Institute for Global Environmental
Strategies (IGES).
Klein Goldewijk, Kees, Arthur Beusen, Gerard van Drecht, and Martine de Vos.
2011. “The HYDE 3.1 Spatially Explicit Database of Human-induced Global Landuse Change over the Past 12,000 Years.” Global Ecology and Biogeography 20 (1)
(January 1): 73–86. doi:10.1111/j.1466-8238.2010.00587.x.
C2500
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
Klein Goldewijk, Kees, December 2011, personal communication
Peoples MB, Brockwell J, Herridge DF, Alves BJR, Urquiaga S, Boddey RM et al (2008)
Biological nitrogen fixation by food legumes. In: Kharkwal MC (ed) Food legumes for
nutritional security and sustainable agriculture. Proceedings 4th International Food
Legumes Research Conference (IFLRC-IV), New Delhi, India. Indian Society of Genetics and Plant Breeding, New Delhi (in press)
Sheldrick, W. F, J. K Syers, and J. Lingard. 2002. “A Conceptual Model for Conducting
Nutrient Audits at National, Regional, and Global Scales.” Nutrient Cycling in Agroecosystems 62 (1): 61–72.
BGD
9, C2463–C2503, 2012
Interactive
Comment
Sitch, S., B. Smith, I. Prentice, A. Arneth, A. Bondeau, W. Cramer, J. Kaplan, et al.
2003. “Evaluation of Ecosystem Dynamics, Plant Geography and Terrestrial Carbon
Cycling in the LPJ Dynamic Global Vegetation Model.” Global Change Biology 9: 161–
185.
Sivakumar, M. V. K., H. M. Taylor, and R. H. Shaw. 1977. “Top and Root Relations of
Field-grown Soybeans.” Agronomy Journal 69 (3): 470–473.
Smil, V. 2002. “Nitrogen and Food Production: Proteins for Human Diets.” Ambio 31
(2): 126–131.
Smil, Vaclav. 1999. “Nitrogen in Crop Production: An Account of Global Flows.” Global
Chemical Cycles, 13: 647–662.
Weindl, Isabelle, Hermann Lotze-Campen, Alexander Popp, Benjamin Leon Bodirsky,
and Susanne Rolinski. 2010. “Impact of Livestock Feeding Technologies on Global
Greenhouse Gas Emissions.” In Stuttgart, Germany,.
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Wirsenius, S. 2000. “Human Use of Land and Organic Materials”. Göteborg, Sweden:
Chalmers University of Technology and Gäteborg University.
Please also note the supplement to this comment:
C2501
Discussion Paper
http://www.biogeosciences-discuss.net/9/C2463/2012/bgd-9-C2463-2012supplement.zip
BGD
9, C2463–C2503, 2012
Interactive comment on Biogeosciences Discuss., 9, 2755, 2012.
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
C2502
BGD
9, C2463–C2503, 2012
Interactive
Comment
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
Discussion Paper
Fig. 1. Table A7: Cropland expansion and release of Nr from soil organic matter loss.
C2503